Incidental venous thromboembolism in oncology patients


Alok A. Khorana, MD, Department of Medicine, James P. Wilmot Cancer Center, University of Rochester, 601 Elmwood Ave, Box 704, Rochester, NY 14642, USA.
Tel.: +1 585 275 4797; fax: +1 585 273 1042.


Khorana AA, O’Connell C, Agnelli G, Liebman HA, Lee AYY, On Behalf of the Subcommittee on Hemostasis and Malignancy of the SSC of the ISTH. Incidental venous thromboembolism in oncology patients. J Thromb Haemost 2012; 10: 2602–4.

Cancer-associated thromboembolism is a frequent clinical complication, particularly in the context of antineoplastic therapy. The introduction of multidetector computed tomography (MDCT) scanner technology has led increasingly to the identification of venous thromboembolism (VTE) on scans ordered primarily for staging or restaging of malignancy. Such incidentally discovered VTEs are variously referred to in the literature as ‘incidental’, ‘asymptomatic’, ‘unexpected’ or ‘unsuspected’ VTE. There are emerging data describing the prevalence, prognostic implications and treatment options for such incidentally discovered VTE events.

The overall prevalence of incidental VTE among cancer patients undergoing routine staging MDCT scans varies with the study population. Studies have reported incidental findings of both pulmonary embolism (PE) and deep vein thrombosis (DVT) identified on computed tomography (CT) scans of the abdomen and pelvis, as well as splanchnic or visceral vein thrombi. In a recent large systematic review and meta-analysis of 12 studies including > 10 000 patients, cancer patients had a weighted mean prevalence of incidental PE of 3.1% (95% confidence interval [CI] 2.2–4.1) [1]. In a large retrospective cohort analysis of scans in cancer patients, Douma et al. [2] reported a prevalence of an incidental abdominal DVT of 1.1% (95% CI 0.6–2.0), which was similar to the prevalence of a PE or lower-extremity DVT (1.3%, 95% CI 0.7–2.3). A higher prevalence of abdominal DVT (2.50%, 95% CI 1.78–3.48) was observed by Ageno et al. [3] in a subgroup of cancer patients. In higher-risk cancer patients, the prevalence can be much higher. Singh et al. [4] evaluated consecutive gastrointestinal malignancy patients undergoing routine staging scans, and found that 7.3% had unsuspected deep vein and visceral venous clots, with incidental PE being found in 2.3% of patients. The most recent reports suggest that a high proportion of cancer-associated VTEs are incidentally discovered. In a retrospective cohort analysis by Moore et al. [5], 44% of all thromboembolic events were incidental. In the cohort study of Singh et al. [4] discussed earlier, 50% of DVTs and > 35% of PEs were incidentally discovered.

The consequences of incidentally diagnosed VTE appear not to differ significantly from those associated with suspected VTE. In a recent analysis, rates of VTE recurrence, bleeding and mortality were similar in cancer patients with incidental VTE as in cancer patients with symptomatic VTE [6]. In a case–control study, cancer patients with incidental VTE had significantly worse survival (hazard ratio [HR] 1.51, 95% CI 1.01–2.27, P = 0.048) than matched cancer patients without VTE [7]. In patients with pancreatic cancer, DVT (HR 25, 95% CI 10–63, P < 0.0001), PE (HR 8.9, 95% CI 2.5–31.7, P = 0.007) and incidental visceral events (HR 2.6, 95% CI 1.6–4.2, P = 0.0001) were all independently associated with mortality [8]. Furthermore, the pulmonary distribution of incidental emboli is no different than of symptomatic emboli, with nearly half being in major pulmonary vessels [9].

Thus, incidentally discovered PE and DVT represents a prevalent clinical problem among cancer patients, and contributes significantly to the burden of cancer-associated VTE. However, there remains considerable confusion among investigators with regard to the appropriate nomenclature and definitions of such events, and whether such events should be included in endpoints of prospectively designed cohort and interventional studies. In addition, there is confusion among clinicians with regard to appropriate reporting and treatment choices for individual patients. The objective of this ISTH statement is to propose a standardization of the nomenclature, definition and reporting of incidentally discovered VTE. Such standardization is necessary to ensure consistency in reporting and to facilitate summary of results across studies.


  • 1 Nomenclature: We recommend the use of the term ‘incidental’ and recommend against the use of the term ‘asymptomatic’. Chart reviews of patients incidentally diagnosed with VTE suggest that many of them are, in fact, symptomatic, with symptoms possibly being attributed to the underlying malignancy rather than to VTE [10]. Thus, in the majority of patients, these events are not truly asymptomatic. ‘Unsuspected’ VTE may also be used; however, the majority of reports in this field have used the term ‘incidental’, and this is concordant with the use of the term in other such diagnoses in medicine (e.g. incidental lung nodules or incidental adrenal adenomas). We recommend using the term ‘symptomatic’ VTE to indicate cases where patients were investigated specifically for VTE on the basis of signs and symptoms.
  • 2  Radiologic issues: Radiology-related factors, including slice thickness and the sensitivity of the reader, impact on incidental VTE prevalence. In a meta-analysis of studies including patients with and without cancer, Dentali et al. reported a weighted mean prevalence of incidental PE of 3% with CT scans utilizing a slice thickness of < 5 mm, as opposed to 2% with scans utilizing a slice thickness of at least 5 mm [1]. Browne et al. reported that CT scans utilizing a slice thickness of 5 mm missed 39% of PEs identified with the use of thinner slices in CT pulmonary angiography scans [11]. Engelke evaluated the accuracy of PE diagnosis in patients with and without suspicious symptoms, and found that PE was more likely to be missed when clinical suspicion was low (89.5% false-negative rate among routine staging scans for esophageal cancer) and when the clot burden was low [12]. Indeed, few data are available regarding the prevalence and accuracy of isolated subsegmental pulmonary emboli, the detection of which may be heavily influenced by the expertise and threshold of suspicion among individual radiologists [13]. Isolated subsegmental PE constituted 6% of incidental PEs identified in the series reported by Gladish et al. and 24% in the series reported by O’Connell et al. [10,14]. We urge cancer providers in multidisciplinary teams to discuss these issues with their radiology colleagues, and to ensure awareness of incidental VTE within their institutions, given the known clinical consequences of such diagnoses.
  • 3  Reporting of events: We recommend, where possible, that retrospective and prospective studies (including clinical trials) separately report rates or proportions of VTE that are incidental or symptomatic. We strongly encourage identification of the regional vessel involved (e.g. ‘portal vein thrombus’ rather than ‘abdominal vein DVT’), as well as the provision of known information regarding the extent of the burden (e.g. isolated subsegmental PE). Where known, investigators should also report radiologic factors such as slice thickness, which can influence the sensitivity of VTE detection.
  • 4  Reporting of outcomes: We urge, where possible, that retrospective and prospective studies (including clinical trials) report outcomes, including recurrent VTE, other synchronous or metachronous VTE, failure rates on anticoagulation, and survival, separately for incidental and symptomatic VTE.
  • 5  Study endpoints: We leave to individual research teams the decision regarding whether or not to include incidental VTE as part of primary endpoints for studies of prophylaxis or treatment of VTE in cancer. We note, however, that the preponderance of evidence suggests that outcomes for patients with incidental PE are not significantly different from those for patients with symptomatic PE (with the possible exception of isolated subsegmental PE, for which additional imaging with lower-extremity compression ultrasonography may be beneficial [10]), and that most clinicians do, in fact, anticoagulate patients with incidental PE [6].

Disclosure of conflict of interests

A. A. Khorana is supported by grants from the National Cancer Institute K23 CA120587, the National Heart, Lung and Blood Institute 1R01HL095109-01, and the V Foundation. C. O’Connell receives grant support from the Whittier Foundation and Stand Up to Cancer. The other authors state that they have no conflict of interest.